Mapbox is preparing to launch a new organised editing effort named Mapbox Trees, starting in California, USA.
The main goal of this project is to map trees (natural=tree) and tree rows (natural=tree_row) to improve the completeness of natural features and enhance the visual quality of green coverage on the map.
Geographical Scope: USA (starting focus on California)
Start Date: We plan to begin editing in two weeks, on October 5, 2026
Pilot Phase Note: During our initial methodology and imagery validation phase, a small pilot batch (~1.5k trees) was mapped under our general #mapbox_editing tag. To ensure clear project tracking and isolation, all ongoing and future mapping for this effort is moving exclusively to #mapbox_trees.
Sources & Tools: We will be using standard aerial orthophotos (Bing, Mapbox, Esri) alongside compliant street-level imagery (Mapillary, Bing Streetside) within the iD editor.
Team: All participating team members have completed internal training sessions alongside standard OSM guidelines. A full list of active mappers is maintained on our GitHub page: OSM Data Team Members · mapbox/mapping Wiki · GitHub
We welcome feedback, suggestions, and local guidance from the community to ensure our edits align with local mapping patterns. Please feel free to share your thoughts here, on the GitHub page, or directly on changesets with our project hashtag.
Hi there! I’m a little surprised that tree mapping is what Mapbox is spending organizing effort on, but more power to you! I can’t seem to find the pilot mapping project, can you link to some of those changesets?
On your github, you say you will use leaf_type or leaf_cycle. Can you detail how you plan to assess those from aerial imagery?
California in particular is an interesting case to start with, as it has a wide variety of climate types and environments, from urbanized areas to deserts to scrubland to thick forests. Can you clarify how you’ll be choosing which areas to map?
Hi everyone, thanks for the feedback and questions!
Can you clarify how you’ll be choosing which areas to map? / Can you be a little more clear on subdivisions - e.g. focusing on urban trees…
Our focus is strictly on urbanized areas, primarily mapping trees along major streets, highways and inside popular territories. We will not be mapping trees in national parks, thick forests, scrublands, or wilderness areas.
On your github, you say you will use leaf_type or leaf_cycle. Can you detail how you plan to assess those from aerial imagery?
We do not plan to apply leaf_type or leaf_cycle universally. These tags will only be added when they can be verified with high confidence — primarily using street-level imagery (such as Mapillary or Streetside) for distinct, easily identifiable species like palm trees or evergreens/conifers. If the tree type cannot be determined with certainty, we only apply the basic natural=tree tag.
I can’t seem to find the pilot mapping project, can you link to some of those changesets?
Here are a few examples of those changesets for reference:
Thanks, these are helpful and reasonable responses! My only remaining note is that I see in the changesets you provide (which look good), the mapper used a couple of tags not documented in the workflow, like taxon=Arecaceae for palm trees. The usage itself looks correct to me, but it’d be helpful to add that you’ll be using those tags to the documentation as well.
